Concept for a Self-correcting Sheet Metal Bending Operation · 2017-01-28 · Concept for a self-correcting sheet metal bending operation U. Damerowa*, D. Tabakajewa, M. Borzykhb,

Procedia Technology 15 ( 2014 ) 439 – 446

2212-0173 © 2014 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).Peer-review under responsibility of the Organizing Committee of SysInt 2014.doi: 10.1016/j.protcy.2014.09.003

Available online at www.sciencedirect.com

ScienceDirect

2nd International Conference on System-Integrated Intelligence: Challenges for Product and Production Engineering

Concept for a self-correcting sheet metal bending operation

U. Damerowa*, D. Tabakajewa, M. Borzykhb,W. Schaermannb, W. Homberga, A. Trächtlerb

aForming and Machining Technology (LUF), University of Paderborn, Warburger Str. 100, 33098 Paderborn, GermanybFraunhofer Institute for Production Technology (IPT) – Project group Mechatronic Systems Design,

Zukunftsmeile 1, 33102 Paderborn, Germany

Abstract

Geometrical deviations can appear in the production of plug contacts used in electrical connection technology and in fittings for the furniture industry. The reasons for this can be a variation in the properties of the semi-finished product, or wear phenomena on the forming machine itself or on the bending tools. When geometrical deviations appear, the process parameters normally have to be adjusted manually. Finding the most appropriate process parameters is currently done manually and is thus very time consuming. In order to reduce the scrap rate and the setup time for production scenarios, a concept for self-correcting bending operations is being developed using the V-model of the VDI guideline 2206. In this case, the V-model will make it possible to set up a self-correcting control strategy consisting of a closed-loop control approach, measurement devices and actuators. Having implemented these components in the forming machines, it will be possible to recognize geometrical deviations automatically and to take corrective action during production, aiming at a reduction of the scrap rate and setup-time.

© 2014 The Authors. Published by Elsevier Ltd.Peer-review under responsibility of the Organizing Committee of SysInt 2014.

Keywords: self-correcting process; punch bending; continuous bending process; design methodology

* Corresponding author. Tel.: +49-5251-60-5349; fax: +49-5251-60-5342.E-mail address: [email protected]

© 2014 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).Peer-review under responsibility of the Organizing Committee of SysInt 2014.

440 U. Damerow et al. / Procedia Technology 15 ( 2014 ) 439 – 446

1. Introduction

Complex metal parts concealed in everyday devices, such as in plug contacts for electrical connection technology and pull-out rails for the furniture industry, are produced by forming processes. The challenges involved in the production of these metal parts lie in the time-intensive setup of the bending processes and compliance with the product requirements during production. Today, when undesirable geometrical deviations appear, the machine operators have to set new process parameters on the basis of their experience. These targeted interventions are only possible when the production run has been stopped, and hence this procedure is very time consuming, especially when it has to be performed more than once. Besides that, the frequent production of out-of-tolerance parts leads to high scrap rates.

New techniques for controlling manufacturing processes are necessary to ensure that the requirements on thefunctionality and quality of the produced parts are met and that the setup time is shortened. The approach presented here involves the development of self-correcting forming processes. The rapid and successful development of new,intelligent manufacturing processes is supported by the use of the V-model [1]. This paper presents a systematic approach for self-correcting forming processes which focuses on the process design with the help of the V-modeland hence not on a detailed description of the individual steps.

2. Intelligent forming process

The aim of this work is to develop intelligent forming processes which will be able to react adaptively to process properties as well as to variations in the raw material. A self-correcting control strategy is thus implemented in the processes. The current process consists mainly of machine-drive and forming tools, which are adjusted manually. The workpieces are then produced from the semi-finished product and measured after the process. If geometrical deviations are detected in a workpiece, new process parameters will be set manually.

To avoid the complex and time-consuming setup of the process, the current process is extended by self-Xcomponents. These are actuators, measurement systems and a closed-loop control (Fig. 1). The measurement system obtains the current information from the process on-line. After the corrective action has been calculated, the actuators modify the process parameters. This leads to a so-called on-line process control system.

Fig. 1. Operating structure of the process.

The development of new self-X components is highly complex and is thus performed through the consistent application of a design methodology that involves primarily model-based design and analysis. The correcting strategy with self-X components is being implemented in prototype plants in order to successfully achieve an intelligent forming process.


3. V-model for intelligent forming processes

Only a few design methodologies currently exist for designing metal forming processes, and these are mostly focused on specific individual cases. In the field of mechatronics, by contrast, a general interdisciplinary design methodology known as V-model is available. A systematic approach is often chosen for the development of self-Xtechnologies. This is an iterative approach, taking in several cycles if necessary (Fig. 2). Normally, the V-model is applied for the development of new products in the field of mechatronic systems. It can also be applied to the design of forming processes, however, as is demonstrated in this paper. In this case, no new forming process is to be developed. Instead, the V-model is being applied to existing processes in order to extend them with self-Xcomponents.

As with the design of mechatronic systems, the model-based approach has priority here. The V-model starts with a process analysis, followed by modeling and model analysis (Fig. 2). The simulation models that are generated undergo constant improvement on the basis of process analysis and other tasks conducted during the overall development phase. Afterwards, suitable actuators and sensors are developed and a control unit designed. These are also being mapped in the model and, after this, prototypically tested on test facilities. Finally, all the components are to be integrated in the prototype plant. The V-model is presented in what follows, taking two different application examples. For the field of electrical connection technology the forming process that is used is punch-bending, and for producing furniture fittings, the continuous bending process.

Fig. 2. V-model used for the design of intelligent forming processes according to [1].

4. Example processes

The development of self-correcting forming processes is demonstrated by means of two application examples, apunch-bending and a continuous bending process. On the basis of these processes, typical requirements for self-Xtechnologies are described, and the method employed for implementing the self-correction in near-serial pilot applications validates the proposed design approach.

The punch-bending process sets out to implement the production characteristics of miniaturized connections and contacts in production processes running at up to 800 strokes/minute. Fig. 3 shows a typical production process for acontact spring made of a high-strength flat wire. In the first step, the two punches and the dies shape the flat wire. In the next step, the shaped flat wire is bent around the bending core, and the workpiece is finished. Throughout the


process, the workpiece is fixed by a downholder. A slide unit mechanically drives the punches and, consequently,their movement is fixed. With the help of self-X technologies it ought to be possible to implement corrective action in the process so as to guarantee continuous product compliance.

Fig. 3. Production steps for the contact spring.

The continuous bending process is used for producing pull-out rails. As Fig. 4 shows, a metal strip is progressively formed into the desired end cross-section in this process by passing it through pairs of forming rolls (forming stations) positioned one after the other.

Fig. 4. (a) continuous bending process and (b) investigated profile.

The self-correcting should make it possible to automatically recognize the changes in the forming process (such as the thickness of the metal band). Differing from the in [2] described calibration station at the end of the production line needed to inter alia counter the spring back, the self-correcting should automatically correct defined critical parameters on the individual stage of the continuous bending process to maintain product quality. This should then ensure that the continuous bending process can robustly withstand material influences and permit the use of metal bands with wide tolerances for the process without endangering product quality. In addition, both the material costs and the scrap rate for production can be reduced.

5. Steps of the V-model

The steps of the V-model will now be presented. In addition, a brief overview is included of the specific application of the steps to the example processes employed.

5.1. Analyzing the processes

The first step in the V-model covers the process analysis. For an in-depth understanding of both the punch bending and the continuous bending process, it is necessary to analyze the process design, tool design, and the


behavior of the machines. When analyzing the processes, all the factors relative to accuracy as well as the influences on a future process control system ought to be identified. Both processes were thus analyzed in series production, and simulation models and testing facilities were set up.

When analyzing the punch bending process, the static and dynamic behavior of the cam disc machine used in series production is first investigated. By using simple multi body simulation models (MBS), it is possible to obtaingeneral information on the velocities and forces occurring in the process [3]. This enables the required measurement devices, such as distance and force sensors, to be set. The analyses are aimed at obtaining much more precise information on the machine behavior. This includes the speed of the machine, the forces and the travel of a single axis of the machine as well as the dimensional accuracy of the bending parts produced. In addition, influences such as the heating of the machine during its operation, machine-specific characteristics and the possibilities for corrective action through changing the punch movement, for example, were examined. The analyses proved successful and made it possible to define requirements for the future actuator such as the location within the process, and the necessary maximum force and travel.

In the case of continuous bending, tests were performed on a serial production system. These were initially used to gain a general overview of the process. It has been shown that there are interactions between the individual forming stations, but no direct relationship could be established between a manipulated variable (e.g. the adjusted height of the forming roll) and a characteristic property of the metal strip. It has, however, been observed that varying the thickness of the metal strip leads to a corresponding change in the forces acting on the upper forming rolls. Furthermore, a higher production rate also led to an increase in the force amplitude. The results that were obtained could be used, for example, to plan the requisite self-correction actuators for the maximum force occurring in the process. In addition, the results are essential for parameterizing the simulation models that are presented in the following section.

5.2. Simulative procedure

The data obtained from the process analysis are used for modeling and simulating the bending processes. A number of modeling methods for metal bending processes have been developed. Heller et al., for example,developed a mathematical model for a fast semi-analytical simulation of the air bending process of thin as well as of thick sheets [4] and Ridane enhanced this approach by regarding the elastic machine behavior [5]. To investigate and consider the spring back Panthi set up a total-elastic–incremental-plastic algorithm, for large deformation and large rotational problems [6]. In this case, the simulation will serve to identify both the relevant process parameters and the appropriate engagement positions for corrective action. Then, based on these simulation models, the necessary corrective strategies can be designed and analyzed. This makes it possible to significantly cut back onexpensive and time-consuming experimental tests on the test facilities.

The forming of the punch bending and continuous bending parts was simulated with the aid of the finite element method (FEM) and also with MBS models. The use of FE-models, with ANSYS for the punch bending and alsoUBECO Profil and ABAQUS for the continuous bending, permits much more precise simulation of the plastic behavior than an MBS model built up with RecurDyn. An MBS model, however, has a major advantage when it comes to designing and testing a control strategy for corrective action in the process models [7]. It is also possible to simulate the dynamic behavior of the production machines, but the simulation of the plastic behavior of the bending parts is rather imprecisely computed by comparison to FE models.

The simulation with the model is used to identify relevant process parameters and their interrelationships, e.g. the influence of the tool position and semi-finished geometry on the part geometry. The result of the model-based analysis is the correction concept, including the definition of the measurement and the control variables.

Despite its lower level of accuracy for the plastic behavior of the bending process, the focus of the MBS model is on the development and testing of the correction strategy. The correction strategy is developed on the basis of the simulation results and the process analyses. The process model is extended by the control part. Employing idealized actuators and sensors, initial knowledge about the controlled process is obtained and evaluated. The advantage here is that an on-line process control can be tested without implementing the necessary self-X components in hardwareform.


5.3. Measurement and actuators

After analyzing the processes, the next step in the V-model (Fig. 2) is to develop the measurement devices and actuators. To capture typical characteristics of the part produced, such as a geometrical reference dimension, it is necessary to select appropriate sensors. To implement corrective action in the process, e.g. for adjusting ageometrical reference dimension, special actuators need to be developed. To define significant measuring and intervention points, use is made of the results of process analysis and simulation. Sensors will measure the current state of the process and provide the information to the control unit. The control unit gives the actuators the information required to adjust the desired process variable. In this case, the knowledge of other parameters like tensile strength or elongation is not important, because it is planned to correct the deviation from the setpoint independently from these parameters.

The measurement technology in the punch bending process is used both for monitoring the bending process andfor on-line process control. It is divided into two groups: one is for the acquisition of process variables such as the punch position and the bending force, while the other is for the collection of the relevant parameters of semi-finished products or articles. The challenges for the measurement are precision, dynamics and integration into the existing tool. Laser sensors are thus selected for position measurement, and piezo sensors for force detection. For detecting geometric features, optical measuring devices such as camera systems can be used. A check should, however, be conducted to see whether the existing market products are suitable for meeting the requirements. During the development of actuators for the punch bending process, the challenges are as follows: short movement, high dynamic, self-locking and integration into existing tools. There are a lot of potential solutions for the actuatorwhich are soon to be tested on the test facility (Chap. 5.5).

The challenge in the development of actuators for the continuous bending process is to produce high linear forces within a small space. In this case, a boundary condition is that the currently used manual forming station has to be expanded by the controllable elements. This has the advantage that most existing conventional systems can be extended by actuators for little outlay. In the case of the sensor, what counts is that measurements must be able to be captured online during the production process. Furthermore, the geometry ought to be determined during the downtime of the metal strip to ensure that the measurement is made on a defined cross-section. Because of the accuracy requirements for the profiles, the measured quantities must be of an appropriate quality, which means that not every measurement method can be employed. A market analysis has revealed that there is no system available yet meeting all the required criteria, so that own developments are necessary.

In both cases – punch bending and continuous bending – the sensors and actuators will be tested and optimized on the test facilities (Chap. 5.5). This is to be followed by integration in the prototyping system.

5.4. Control strategy

Using the V-model (Fig. 2), the self-correcting strategy will now be designed. Its main component is a closed-loop control (Fig. 1) for the on-line process control. At the same time, the design of the controller constitutes a particular challenge, since an incorrect interpretation can lead to uncontrolled behavior [8].

In the punch bending process, the challenge for the correction strategy lies in defining the variables from the forming process which are of relevance for online process control [9]. The punch motions are used as actuating variables. The control system itself consists of a feedback controller and disturbance compensator. The design of thecontroller parameters is performed on the basis of models using optimization algorithms. The first approach involvesthe use of a discrete PI controller. The controlled variable is a relevant geometrical characteristic of the workpiece for the opening dimension, for example (Fig. 3). Hence, in the case of deviations of the actual geometry from the desired geometry, the corresponding axes will be adjusted. If the opening dimension of the actual part is larger than the desired size, one possibility is increasing the punch movement to adjust the opening dimension of the next part.

Due to the high number of degrees of freedom in the continuous bending process, controller design is a challenging task. It is difficult to find a solution for the control concept as there are a number of mutually influencing interactions between the forming stations. In addition to this, a number of input variables have to be processed, and multiple output variables have to be emitted. Currently, it is planned to record profile characteristics such as the opening dimension (Fig. 4 b) at different positions in the manufacturing process and to use them as


inputs for the control unit. As output variables, the upper roll positions of selected forming stations are used to vary the opening dimension. In this case, an appropriate option is the model-based design of a multivariable control [10]. Hereby the control strategy is first designed and tested on the model and then implemented on the real prototype plant.

5.5. Test facilities

Since most of the constructions used in this paper are new developments, they are examined on test facilities before being implemented in the prototype system. Test facilities have been designed and constructed for both processes. This makes it possible to investigate the behavior of the forming processes under realistic conditionswhile still benefitting from a laboratory situation.

Fig. 5. Setup of the test facility for (a) the punch bending and (b) the continuous bending process.

The test facility for punch bending consists of a bending tool, an original cam slide unit from series productionthat is driven by a motor, two laser displacement sensors and a force sensor (Fig. 5 a). The slide unit moves a punch within the bending tool, and the punch movement and force can be measured. The bending tool permits manual adjustment of the thickness of the flat wire used, which can then be checked with the second laser displacement sensor. The test facility can also be used to examine different actuator concepts to intervene in the mechanical cam driven process and to investigate different strategies for a closed-looped control for several different production scenarios.

In the case of continuous bending process, a test stand was constructed (Fig. 5 b) to investigate the actuators for adjusting the upper forming roll. The actuator has a motor at the top, whose torque is stepped down via a transmission. The rotation is then converted by a screw drive into a translational movement, whereby the upper bearing block is moved. This is used for supporting the upper forming roll in the real structure, but this is not the


subject of the investigation. In order to simulate the process forces, a force-controlled hydraulic cylinder is adjusted in relation to the bearing block. The force that acts is captured via a force sensor above the bearing block. Using this test stand, it is possible to impose previously determined process forces, for example. The object of the study included achieving correct dimensions on the mechanical and electrical components. Further study variables were the impact of the backlash in the components on the positioning accuracy of the bearing block and its position-retention accuracy.

The test facilities for both processes make it possible to work independently of the machines used for series production which have high utilization rates. In this way, the measurement devices and actuator can be tested and optimized while having the control strategy that has been developed implemented in an industrial control system. With the results obtained it will soon be possible to transfer the control strategies, including all the components, to a prototype in a series production machine.

6. Conclusion

Up to now both punch-bending and continuous bending are mechanical processes that have been adjusted manually on the basis of the machine operator's experience. Unavoidable variations in the process parameters have led to long setup times and high scrap rates. To remain competitive it is thus necessary to reduce the setup time and scrap rate and ensure a consistent workpiece quality throughout the process. The first concepts for a self-correcting punch-bending and continuous bending process have been developed on the basis of the V-model. This methodology permits the interdisciplinary development of self-X components and their implementation in forming processes. With self-correcting processes it will be possible to drastically reduce the setup time and to consistently keep the workpieces within the tolerances.

Acknowledgements

This research and development project is being funded by the German Federal Ministry of Education and Research (BMBF) within the Leading-Edge Cluster “Intelligent Technical Systems OstWestfalenLippe” (i.e. OWL) and managed by the Project Management Agency Karlsruhe (PTKA). The authors are responsible for the contents of this publication. We gratefully acknowledge the collaborative work and support of our project partners Weidmüller Interface GmbH & Co. KG and Hettich Holding GmbH & Co. oHG.

References

[1] VDI guideline. Design methodology for mechtronic systems (VDI 2206). Berlin: Beuth Verlag; 2004[2] Henkelmann M. Entwicklung einer innovativen Kalibrierstrecke zur Erhöhung der Profilgenauigkeit bei der Verarbeitung von höher- und

höchstfesten Stählen. PhD-Thesis, University of Darmstadt; 2008.[3] Gattringer H, Gerstmayr J. Multibody System Dynamics, Robotics and Control. Berlin: Springer; 2010.[4] Heller B, Kleiner M. Semi-analytical process modelling and simulation of air bending. The Journal of Strain Analysis for Engineering Design

2006; 41(1):57-80.[5] Ridane N, Jaksic D, Kleiner M, Heller B. Enhanced semi-analytical process simulation of air bending. Advanced Materials Research 2005; 6-

8:729-736.[6] Panthi SK, Ramakrishnan N, Ahmed M, Singh SS, Goel MD. Finite Element Analysis of sheet metal bending process to predict the

springback. Materials & Design 2010;31(2):657-662.[7] Borzykh M, Damerow U, Henke C, Trächtler A, Homberg W. Model-Based Design of Self-Correcting Strategy for a Punch Bending

Machine. Information Control Problems in Manufacturing 2012;14(1):1551-1555.[8] Föllinger O. Regelungstechnik - Einführung in die Methoden und ihre Anwendung. Berlin: VDE Verlag GmbH; 2013.[9] Damerow U, Borzykh M, Homberg W, Trächtler A. A self-correcting approach for the bending of metal parts. Key Engineering Materials

2012;504-506:907-912.[10] Skogestad S, Postlethwaite I. Multivariable Feedback Control - Analysis and Design. New York: Wiley-Interscience; 2005.

Concept for a Self-correcting Sheet Metal Bending Operation · 2017-01-28 · Concept for a self-correcting sheet metal bending operation U. Damerowa*, D. Tabakajewa, M. Borzykhb,

Documents